Outdoor Minisplit Unit Freezing up in Temperatures Above Freezing
I’ve used cheaper ducted mini split units ( Pioneer, Gree ) in my house with great success so far, but now as we are renovating my son’s house we decided to splurge and go with a Fujitsu 18RLFCD. We are in DE, climate zone 4.
We installed the unit yesterday, fired it up and let it run overnight to warm up the just closed in house a little ( no insulation yet, but pretty tight construction).
In the morning the outside conditions were low 30s with 85% humidity, the temperature inside was only 48 F and the ducts were blowing barely lukewarm air.
I’ve checked the outside unit and it was like a big snowball, the radiator fins are totaly covered with ice.
To compare, the fins on the Pioneer unit in my house are completely dry in the same weather.
Anybody have an idea why the Fujitsu is freezing up?
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Replies
Sounds like a major malfunction. Is--or was-- the outdoor fan running, and expelling lots of cold air?
Looking at the photo if the entire coil looks like the photo the units sensors should tell the computer it is time for a defrost cycle.
If some part of the coil is still clear the unit will defrost soon when that part gets covered.
If it just keeps getting worse without any defrost cycles you may have a faulty sensor.
If the unit seems to be defrosts more often than normal that could be a symptom of the unit being under charged and or having a leak.
Walta
Defrost control is a very hard problem, manufactures have to balance efficiency (as few defrost cycles as possible) and keeping the coil clear. This is especially hard in near freezing weather with high outdoor humidity.
I've had one that had no problems any other time frost up in similar conditions, the unit kept going into defrost mode but was just not able to clear the coil. Ended up running in cooling for a bit which did the trick.
Looks like pretty light frost on there, wouldn't worry too much about it. If the unit is starting to turn to an ice ball, you have some other issues.
The outdoor fan is running but not at full speed. The lineset is warm, but not hot. I gess the unit is turning down sensing the reduced airflow., Measured the charge, looks correct. No leaks either.
I haven't had much time to watch the unit yet, but never seen or heard it to do a defrost cycle. In the morning it had about an inch of ice on the bottom of the base pan. I've cleaned it out with a heat gun, but the ice is coming back after the restart. It looks like the weather is just cold and humid enough that the dripping condensation water forms icycles on the drain holes and sooner or later they are getting blocked with ice.
Just frustrating that the cheapo units never had similar problems in any weather.
I have the same unit and the coils on mine will definitely frost up in similar conditions. However, when that happens it runs a quick defrost cycle and within a few minutes the coils are clear again and it's back running. Is it possible that it's just not capable of heating the house without insulation in it? If the unit is forced to work too hard in that kind of weather, that can lead to more frost on the unit. Though if the interior is only set to 48º, you wouldn't think that's too much of a temperature differential at all between inside and out, even without insulation. Though regardless, it should be running defrost cycles when that happens.
The base pan freezing up sounds odd too, given the temperatures. My understanding is under normal conditions, you need a prolonged period of sub-freezing temperatures before that becomes an issue.
You say “no leaks” exactly how can you know that to be a fact?
The only way anyone could stat that as a fact is to have is to pump all of the refrigerant out of the system and into a tank on a scale and weigh what was removed from the system and have it match the factory spec, seems unlikely for a home owner.
Who installed the unit?
I say keep running the system for a few more days and try logging how often it goes into defrost mode.
Walta
I sat down and observed the unit. 39F outside, 75% RH. 50 deg inside, heating set temp at minimum 60F. Not an unimaginable scenario for regular use, and I had fared much worse with different brand units without a hitch during the construction of my house.
After a while the unit stated a defrost cycle. The defrost took 8 minutes, within the first 3 minutes all the ice was gone from the fins. The interior unit started to heat again at 12 minutes. The unit worked at high fan ( remote set to auto fan) until the 48th min from the start of the first defrost cycle and started a defrost again. Basically 25 % of the time it was not heating. That is not a great duty cycle even for a conventional heating system.
The base pan freezing looks like happened because I raised the unit on only 1" legs above the base, and the drips from the drain holes formed ice bridges between the platform and the unit. they got thicker with time and blocked the holes completely. I've raised the unit an other 2 inches and now it is fine.
I understand the arguments about efficiency and too big load etc, but a different brand worked perfectly in similar conditions. The Fujitsu has much finer fins closer to each other than the others, maybe that's why it is more susceptible for freezing.
The Fujitsu plants the bad feeling into the back of your mind that once you let the house cool down too much you will never get warm until spring.
Sounds like the unit is operating properly. Just because it is frosting up more than other units it doesn't mean there is anything wrong it since it clears it without issues.
The 18RLFCD is good for around 24000BTU at these warmer temperatures. Even if the unit is going into defrost 25% of the time, it will still provide more than enough heat if it was sized for 15F outdoor temperatures.
As for the drain issue, it does sound like it was self inflicted.
Depending on which end of DE you are in, you can still see a bit of snow in the winter time, a unit mounted even 3" off the ground is a bad idea. It should be mounted at least a couple of inches above max expected snow fall plus a bit of safety margin for wind blown snow and snow falling off the roof.
To my ear it is starting to sound like your unit is low on refrigerant.
The stats on my heat pump are 6300 hours 1600 defrosts cycles or 4.9 hours of run time between defrost cycles. My defrost cycle time is between 3 and 4 minutes long . Note my heat pump is a conventional VS unit.
Walta
The unit is about 1 foot above ground level, just the distance between the base and the bottom of the unit is the 3" now. I might switch to a wall console once the exterior cladding is done. Just with an unknown unit I did not know what vibration level to expect, that's why the ground mount now. (Basically one reason to pick this unit was the low noise specs)
The install instruction said the unit is precharged to up to 49 ft lineset, our lines are 38 ft long, so no extra charge added.
During the first night (29F) the electrical consumption was hovering around 1.6 KWh. COP at this temp and load should be around 3-3.5 according to ASHP, which is about 18000 BTU.
So accounting for the defrost time, it's running at 100% of capacity. With a 50F indoor temp and just above freezing outside. Sounds like worst case conditions for icing.
I've found that my unit defrosts fairly regularly in the humid/just above freezing conditions. Maybe not quite as often as what you're experiencing, but not too far off really. In practice, it doesn't cause any issues. The efficiency is still great (COP is well over 3 based on my estimates compared to prior years with baseboard heating) and the defrost cycle never takes long enough for it to really affect the indoor temperature. So I just let it do its thing.
It not being able to bring the building up to temperature is the bigger concern. That would probably only be one of two things: 1) there's something wrong with the unit or the installation which doesn't allow it to output the heat it should; or 2) the unit is working correctly, but the heat loss in the uninsulated building is too great for it bring it up to temperature (at least in a reasonable time frame). Either way, that could certainly be causing the excessive frost you're seeing.
Happens to my Fujitsu as well. It hasn't been above freezing here in over a week. We've seen temps drop to -8F and like other parts of the Northeast, we picked up 44" of inches of snow.
The unit keeps on blowing out out warm (not hot) air so we never pay attention to it.
It would be interesting to measure the ice's effect on efficiency.
Outside when it's close to freezing the air often has much higher moisture content than when it's deep cold outside. Also, the coil frosting and pan heater are two separate issues. If you are getting ice bridging you have an install issue. It's too close to the ground or whatever is under it. Ice should be be able to "reach up" to the base pan. If you're trying to thaw the base pan there are some automotive style heaters that may or may not be usable.
I agree that low on charge is a recipe for frozen evaporators, instead of almost flooding the coil and having a pretty (constant coil temp), when a unit is low on charge the refrigerant tends to more immediately flash to gas and create a cold spot that won't necessarily be seen by the unit's coil temp sensor. I would recommend weighing the charge back in. It's also possible that the unit was correctly charged initially but then leaked down.
My Midea and Gree units never get anywhere near that level of frost on the coils like the 2 pictures posted. I'm thinking the refrigerant charge is not right on those units. It looks like there is zero airflow on those coils with that heavy frost.
Have you measured your supply/return delta t? Measure the intake air and output temperature. At full output my Midea and Gree units will supply 120-135 degree air. Just set the unit to heat-high fan and set the temperature to like 82. That will force the unit to run at max output.
Also an electrical monitor will let you see everything that is going on. Here is a snapshot when one of my units was defrosting often yesterday. Its easy to compare the electrical consumption to the neep ashp directory and get the rough output as well.
You can see the inverter speed spike up right before it defrosts. That seems to be part of the defrost algorithm. The software sees the inverter speed increase and the supply temperature not rise. It also uses the outdoor temp sensor and can actually measure the outdoor fan amp draw to signal low airflow on the outdoor unit.
All the manufacturers openly publish their service and technical manuals online. The defrost cycles are explained in it.
https://ashp.neep.org/
I have the same Fujitsu unit but in the 12k size and am seeing similar frost/defrost patterns. The unit will operate in heat mode for about 45 minutes and then run a defrost cycle for about 8 minutes. I don't have an issue with the unit being able to maintain the set temp and delta T between return air temp and supply air temp is about 45 degrees. The pictures are from this morning and the temp is just above freezing and humidity is in the upper 80's.
Its not good to have the unit right on the ground like that. The basepan has holes in it to drain water off when it defrosts. Yours is sitting in snow and runs a high risk of water freezing in the pan and cracking the coil. The end result is all the refrigerant leaks out and the coil is ruined.
It should be mounted on a stand like this. I have been outside when my unit defrosts and water pours out the bottom drain holes. You can see the water spots in my picture below the unit. There is a heating element in the basepan that keeps the pan ice free.
It's not as bad as it looks. The units are sitting on spacer blocks on top of isolation pads on top of a concrete pad. Not to say I should not have shoveled the 12" of snow we got the other day.
My guess would be that the new unit is doing a ton more work than the others. Rather than steady state temp, you are warming up from a colder temp, and also running in an unfinished space.
I saw that you checked the pressure of the charge. If the refrigerant charge is normal, a possible reason is that the mini split is not capable to heat the house under this weather condition --- the size of the evaporator of the outdoor condenser is not large enough.
9RLFCD, 12RLFCD, 18RLFCD uses the almost the same outdoor unit. Fujitsu uses the different firmware/algorithm to control the output performance range. Similarly, 18RGLXD, 24RGLXD, 30RGLXD, 36RGLXD also uses the almost the same outdoor unit and Fujitsu uses the different firmware/algorithm to control the output performance range.
I suggest that you could check the power consumption of the unit. It should use over 1,500W during high frequency operation. If it cannot use such electricity and you are sure that there isn't any leak, the possible conclusion is that the size of the evaporator is too small under this weather condition, which prevents the compressor from entering the high frequency operation. In this is true, you should probably consider 18RGLXD, which uses much bigger/heavier outdoor unit than 18RLFCD.
I had this exact same issue. I had just hung and finished Sheetrock and needed to warm the house up enough to lay primer. It was about this time of year and even same zone. The unit iced up like crazy, but after a few days it got the whole house up to temperature and I finished air sealing, it’s been perfect ever since. In retrospect, I think I underestimated the load in heating the thermal mass of the entire house from freezing to 68. Every board, stick, and sheet in the place is literally freezing!
Mine had a dip switch setting to defrost more often, which I did but have since reverted back to factory setting.
My unit ran for 2 1/2 hours this morning heating the space from 68 to 72 and never went through a defrost cycle. Temps were just above freezing and humidly was high 60's. I did replace the thermostat yesterday, swapped out the standard non-backlit unit that is impossible for me to read with the backlit unit and the thermostat control set to the wired "remote."